Proc. Natl. Acad. Sci. USA Vol. 87, pp. 5459-5463, July 1990 Biochemistry Cloning and expression of a gene encoding an interleukin 3 -like protein: Identification of another member of the receptor gene family (/lymphokine receptor/hemopoietic growth factor/colony-stimulating factor) DANIEL M. GORMAN*, NAOTO ITOH*t, TOSHIO KITAMURA*, JOLANDA SCHREURS*t, SHIN YONEHARA§, ICHIRO YAHARA§, KEN-ICHI ARAI*¶, AND ATSUSHI MIYAJIMA*II *DNAX Research Institute of Molecular and Cellular Biology, 901 California Avenue, Palo Alto, CA 94304; §The Tokyo Metropolitan Institute of Medical Science, Bunkyo-ku, Tokyo 113, Japan; and $Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo 108, Japan Communicated by Joseph L. Goldstein, April 25, 1990

ABSTRACT Using a monoclonal to the interleu- the cloned IL-3 receptor (4). Thus, it is likely that the IL-3 kin 3 (IL-3) receptor (anti-Aic2), we isolated a cDNA (AIC2B) receptor is a multiprotein complex, and the cloned cDNA from a mouse mast cell line which is homologous to the encodes a binding component of the functional receptor. previously characterized gene for the IL-3 receptor (AIC2A). A sequence comparison ofthe extracellular domains ofthe This cDNA encodes a polypeptide of 896 residues cytokine receptors [the IL-2 receptor p chain (7) and the and has 91% amino acid sequence identity with the IL-3 receptors for IL-3 (4), IL-4 (8, 9), IL-6 (10), granulocyte- receptor. A consensus sequence defining an additional cytokine macrophage colony-stimulating factor (GM-CSF) (11), and receptor family is present in this clone. Compared to the (12)] revealed a common motif, establishing AIC2A clone, the AIC2B cDNA encodes a protein with amino the existence ofa distinct receptor gene family (4, 11, 13). The acid substitutions, insertions, and deletions dispersed through- external domains of these receptors have two conserved out the entire protein. Oligonucleotide probes specific for each regions separated by a sequence unique to each individual cDNA hybridized with different genomic fragments, indicating receptor. Interestingly, the external domain of the IL-3 that the AIC2A and AIC2B proteins are encoded by two receptor can be divided into two homologous units, each unit distinct genes. Fibroblasts transfected with the AIC2B cDNA containing this common structural motif(4). In this report we expressed the protein at the cell surface as determined by describe the cloning of a cDNA homologous to the IL-3 binding with the anti-Aic2 antibody but did not bind IL-3 or receptor gene.** The protein encoded by this gene has the other , including IL-2, IL-4, granulocyte-macro- characteristics of the cytokine receptor family and the ligand phage colony-stimulating factor, erythropoietin, and IL-9 (p40) for this protein is unknown. at concentrations between 1 and 10 nM. An S1 nuclease pro- tection assay was used to discriminate between the AIC2A and AIC2B transcripts. We found that the AIC2B gene was coex- MATERIALS AND METHODS pressed with the AIC2A gene. These results suggest a potential Transfection of COS-7 Cells and Binding Assays. COS-7 involvement of AIC2B in cytokine . monkey cells were transfected with 10 yg of cDNA by the DEAE-dextran method or by electroporation at 220 V with Interleukin 3 (IL-3) is a potent hemopoietic growth factor 960 uF in a 0.4-cm cuvette. Three days after transfection, which stimulates multipotential hemopoietic stem cells; it cells were detached by a 10-min incubation with Dulbecco's stimulates the formation of multilineage colonies in vitro and phosphate-buffered saline (PBS) containing 2 mM EDTA and also maintains spleen colony forming units (CFU-S) in vitro. washed with Hanks' balanced saline solution containing In addition, it serves as a growth factor for committed cells, bovine serum albumin at 1 mg/ml and 10 mM Hepes buffer including mast cells, megakaryocytes, eosinophils, erythro- (pH 7.5). Cells were incubated with 1251I-labeled cytokines blasts, pre-B cells, and potentially pre-T cells, and it induces with or without unlabeled cytokines for 2 hr at 4°C. Cell- the differentiation ofosteoclasts (1, 2). Since IL-3 is produced bound radioactivity was determined by centrifugation mainly by activated T cells and is not produced by bone through oil as described previously (6). marrow stromal cells, it may be important for the expansion Genomic Southern Analysis. Chromosomal DNA digested of hemopoietic cells in an inflammatory response rather than with EcoRI was separated on a 1% agarose gel and trans- in constitutive hemopoiesis in the bone marrow (3). ferred to a nylon membrane filter (Schleicher & Schuell). The IL-3 manifests its multiple biological activities through filter was prehybridized at 42°C in a solution containing 6x binding to the high-affinity IL-3 receptor expressed on var- SSC, 5 mM EDTA (pH 8.0), 10 mM Tris HCl (pH 7.5), 5x ious hemopoietic cells. We recently isolated a cDNA encod- Denhardt's solution, 0.1% SDS, denatured calfthymus DNA ing a mouse IL-3-binding protein (4) by using a monoclonal at 100 Mg/ml, and yeast tRNA at 100 pg/ml (lx SSC = 0.3 antibody, anti-Aic2, raised against a multifactor-responsive M NaCl/0.03 M sodium citrate; lx Denhardt's solution = immature mast cell line, IC2 (5). Fibroblasts transfected with 0.02% bovine serum albumin/0.02% Ficoll/0.02% polyvi- the cloned cDNA bound IL-3 with low affinity, as do various nylpyrrolidone). Hybridization was performed at 42°C for 20 IL-3-dependent cell lines that express the low-affinity IL-3 hr in the same solution with an oligonucleotide probe labeled receptor (6). Although evidence suggests that the IL-3 re- ceptor is linked to a protein in normal Abbreviations: IL, interleukin; GM-CSF, granulocyte-macrophage IL-3-responsive hemopoietic cells, neither kinase activity colony-stimulating factor. nor a consensus sequence for tyrosine kinases was found in tPresent address: Osaka Bioscience Institute, Furuedai, Suita 565, Japan. *Present address: Cetus Corporation, Emeryville, CA 94608. The publication costs of this article were defrayed in part by page charge I"To whom reprint requests should be addressed. payment. This article must therefore be hereby marked "advertisement" **The sequence reported in this paper has been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession no. M34397). 5459 Downloaded by guest on September 23, 2021 5460 Biochemistry: Gorman et al. Proc. Natl. Acad. Sci. USA 87 (1990) with [y32P]ATP by the T4 polynucleotide kinase and purified by a Sephadex G-25 NAP column (Pharmacia). The filter was b A washed at 40'C with 6x SSC. Isolation of RNA and S1 Nuclease Protection Assays. Total cellular RNA was isolated by the guanidinium isothiocyanate E method as described previously (14). The S1 nuclease pro- z tection assays were performed according to the method of Berk and Sharp (15). Briefly, the HindIII fragment of the C.) AIC2-19 cDNA or the AIC2-5 cDNA in the pCEV4 expres- sion vector (4), which covers the 5' end ofthe simian virus 40 promoter to the unique HindIII site ofthe cDNA (see Fig. 2), was isolated. The fragment was dephosphorylated by alkaline Fluorescence Intensity phosphatase and the 5' end of the HindIII cleavage site was 4 labeled with [y32P]ATP and T4 polynucleotide kinase to a B specific activity of 106 cpm/Mg of DNA. Five micrograms of total RNA was hybridized with 104 cpm ofeither the A probe x 3 - or the B probe in a 15-1.l reaction mixture containing 80% E (vol/vol) formamide, 0.4 M NaCl, 40 mM Pipes (pH 6.4), 70 a Mg of yeast tRNA, and 1 mM EDTA for 16 hr at 450C. Then, 0 150 ,ul of the S1 solution, which contains S1 nuclease at 0.3

unit/ml, 250 mM NaCl, 30 mM sodium acetate (pH 4.5), and C) 1 mM ZnC12, was added and the mixture was incubated at c'I 37°C for 1 hr. The protected DNA fragments were analyzed by electrophoresis on a 5% polyacrylamide gel. co m - RESULTS COLD IL-3 - Cloning of cDNAs Encoding the IL-3 Receptor and Its mock AIC2-2 A IC2-5 Homologue. From the IL-3-dependent mouse mast cell line FIG. 1. Expression of AIC2-2 and AIC2-5 in COS-7 cells. (A) MC/9 we isolated a cDNA encoding an IL-3 receptor by COS-7 cells transiently transfected with cDNA (a, AIC2-2; b, using the rat IgM anti-IL-3 receptor antibody (anti-Aic2) as AIC2-5) were stained with anti-Aic2 and fluorescein isothiocyanate- described previously (4). During this process, we isolated conjugated antibody to rat IgM. The shaded area shows the staining additional cDNA clones homologous to the clones encoding profile with the anti-Aic2 and the blank area shows the staining the IL-3 receptor. To screen the cDNA library we employed profile with the second-stage antibody alone. (B) COS-7 cells trans- an expression cloning procedure using COS-7 cells as devel- fected with cDNA were used for the IL-3-binding assay. Cells (5 x the cDNA 104) were incubated with 2.75 nM 1251-labeled IL-3 (1251-IL-3) in the oped by Seed and Aruffo (16). Briefly, MC/9 presence (+) or absence (-) of 10 ,uM unlabeled ("cold") IL-3 for 2 library made in the simian virus 40-based mammalian expres- hr was measured. sion vector pCEV4 (4) was introduced into COS-7 cells. at 4°C and cell-bound radioactivity COS-7 cells expressing the Aic-2 antigen were enriched by The AIC2B Gene and Its Predicted Protein Structure. panning using the anti-Aic2. Plasmid DNA was recovered Nucleotide sequencing of the AIC2-5 and AIC2-22 cDNAs from the selected COS-7 cells and subjected to the same revealed a 95% nucleotide sequence identity with AIC2-26, selection procedure. After three cycles of enrichment, indi- including the 5' noncoding region and the 3' noncoding region vidual plasmids were analyzed. Out of 24 plasmids analyzed, sequence (Fig. 2). The protein encoded by the cDNA has a 13 plasmids had the same 1.6-kilobase (kb) cDNA insert 91% amino acid sequence identity with AIC2-26 (Fig. 3). (AIC2-2), 3 plasmids had the same 2.7-kb cDNA insert Nucleotide substitutions, deletions, and insertions were dis- (AIC2-19), and 1 had a 2.8-kb insert (AIC2-5). All these persed throughout the entire coding region, indicating that plasmids expressed the Aic2 antigen in COS-7 cells upon these two types of cDNA were derived from two distinct transient transfection. Interestingly, while COS-7 cells trans- genes rather than created by alternative splicing. To confirm fected with either AIC2-2 or AIC2-19 bound IL-3, COS-7 two distinct cells transfected with the AIC2-5 cDNA did not (Fig. 1). that AIC2A and AIC2B are encoded by genes, Restriction analysis as well as hybridization with the AIC2-2 we performed genomic Southern analysis using oligonucle- cDNA as a probe indicated that AIC2-5 was closely related otide probes specific for each cDNA (Fig. 4). Probe A, which to but not identical with AIC2-2 or AIC2-19. To distinguish is specific to the AIC2A cDNA, hybridized with an 11-kb these two types of cDNA, we designate the cDNA encoding EcoRI fragment, whereas probe B, which is specific to the the IL-3-binding protein as AIC2A and the other cDNA, AIC2B cDNA, hybridized with a 15-kb EcoRI fragment (Fig. encoding a non-IL-3-binding protein, as AIC2B. 4), indicating that the two types of AIC2 cDNA were derived As the AIC2-2 cDNA hybridized only with an RNA of from two different genes. Isolation of the two genes from a approximately 4.5 kb (4), all the plasmids recovered from genomic library confirmed the existence oftwo distinct genes COS-7 cells were partial cDNAs. We therefore isolated larger (unpublished work). cDNA clones by hybridization using the AIC2-2 cDNA insert The amino acid sequence indicated that the AIC2B protein as a probe. We obtained 18 cDNA clones from the original as compared to the AIC2A protein has one additional poten- MC/9 cDNA library. Eight ofthese were AIC2A and 10 were tial N-linked glycosylation site in the external domain and has AIC2B as determined by restriction analysis. The major form 18 extra amino acids in the cytoplasmic domain. The AIC2B of AIC2B cDNAs including AIC2-5 has a long open reading protein retains the unique feature of the IL-3 receptor frame. Comparison with the other clones showed that AIC2-5 (AIC2A protein); i.e., the external domain of the AIC2B lacked 66 amino acids from the C terminus. We therefore protein can be divided into two homologous units sharing constructed a cDNA (AIC2-522) encoding the entire coding 24% identity and 42% similarity and each unit contains the region by recombining AIC2-5 and AIC2-22, which contains common motif of the cytokine receptors such as four con- the C-terminal region, and used this cDNA for further served cysteine residues and the unique WSXWS motiffound biochemical characterization. near the transmembrane domain (4) (Fig. 3). The cytoplasmic Downloaded by guest on September 23, 2021 Biochemistry: Gorman et al. Proc. Natl. Acad. Sci. USA 87 (1990) 5461

GACCTGACTGGGAGATGGTGGGGAGCGCACTACTCTGGCAGAACTMAATGTCATGGGGCACAGATAAATAGGMAAGACCTGCAACTCACTGGCACTTCGAGGCTCCCGAAGGAGGCTGC I 120 start> T 240 GACACCCCCCTGCCCCATAOCTTCCAGTGCAGCCACCAAAATGCCMAAACACCAGCAAATGGCACTCACATGGGGGCTGTGCTACATGGCACTGGTGGCTCTCTGTTGGGGACACGCG T C A 360 GTGACACAGGCAGCGACAMGCTCCCTCTGAAGACTCTGCACTGCTACAATGACTACACCMCCACATCATCTGCACCTGGCGGCACACAGAGGATGCCCACGGGCTAATCAACATGAC A 0 OT A (C 479 .CCTCTATCACCACCTAGAGAAAAMCAGCCAGTGTCCTGTGAOCTCAGTGAGAAACTCATGTGGTCAGAGTGCCCGTCATCCCACCGCTGTGTCCCCAGAAGATGTGTCATCCCCTAT CT) C a TT AT C 597 ACACGAMTCCATCACAAACGAAGACTACTACTCCTTCCGCCCAGATAGTGATCTGGGCATCCAGCTCATGGTCCCACTTGCCCACAATGTGCAGCCACCACTTCCCAAACGCTCAGC T ATOG C TA C A C C C 0 C CC G A CA 717 ATCAGCTCCTCTGAGGATCGMCCTGCTCGAGTGGAGTGTGTCCCTTGOGGATGCCCAGGTCTCCTGCCTTTCATCAAAGGACATAGAGTTTGAGGTGGCTTATAACGCCCTTCAGGAC C G A A A T T 837 TCCTGGGACGATGCCTACAGTCTCCACACTAGCAAATTTCAOGCTAAC C MCGAGCCAAACCTATTCCTACCCAACAGCATCTATGCCCCCCGTGTGCGCACTCGGCTGTACCCGGGTTCAA TG C G HindUII BuaHI 957 AGCTTGTCTOGGCAGACCCAGCAGATOGAOCCCACAGCCTCACTGOGACTCCCAGCCAGGGGCAAGCCCCCAGCCACACMACCTTCAATGCTTCTTTGATGGGATCCAGTCCCTCCACTGC T T 1077 TCCTGGGaAGTGTGGACCCAGACGACTGCCTCTGMCCMGGCXCTTTCTATCGCCCCAGCCCTGCTACCTCCGGAGCAGAAMTGCTCTCCCGTGGTCGAAGAGCCG(CCG)GGGGCCAGT C (...)A C 1197 GTCTACACCCGCTACCATTGCAGTCTACCTGTGCCTGAGCCCAGTGCACACAGCCAGTACACAGTCTCTGTTAAGCACCTGGAACAAGGGAAGTTCATCATGAGCTATAACCACATCCAGGC T 1317 ATGGAGCCTCCAACCCTCAACCTACCAAGACAGAGACAOCTACAGCCTGCATTGGGAAACTCAGAAGATGCCCTTA)CTCATTCATTGAGCACACATTTCCAGGTCCAGTACAAGMAAGAA A T A A AC(... )AA A T T G 1437 TCGGACAGCTGGGAGGACAGCAAGACAGAGAACCTAGATCGAGCCCATAGCATGGACCTCTCCCAGCTGGACCCAGACACCTCATACTGCGCCACGGTGAGGGTCAAGCCCATCTCTAAC A G A A 0 T A GC 0 G 1557 TACGATGGGATCTOGAGCAAGTGGAGCGAAGAGTACACTTGGAAGACTGACTGGGTGATGCCCACGCTGTGGATAGTCCTCATCCTGGTCTTTCTCATCCTCACCTTGCTCCTGATCCTT C 0 A T C T OCT C 1677 CC~GCMGCTGTGTCTCTGTATACAGGACGTACAGAAMGTGGACA C T A 0G A WATCCCC AACCCCAGCCAAGACCTCCTGTTCCA GGAT GGAGGTAAAGGTCTCTGGCCTCCTGGCAGCATG 1797 GCAGCCTTCGCCACTAAGAACCCCGCTCTCCAGGGGCCACAGAGCAGGCTTCTTGCTGAGCAACAGGGGCAGTCATATGCACATTTGGAGCAACAAMCGTGTCACCTCTCACTATAGAG 0 T A 1917 ACCCTAATATAATTCGAGTTCCACCATCCGGGCCTGATACAACCCCAGCTGCCTCATCCGATCCACAGAGCAACTTCCCATGTTCAGTAGAGCACCM ...CTCCTMCACACCT A A T (TTC) T 0 C C FIG. 2. Nucleotide sequence 2034 ACGAGCAATTACCCAGCMGACTTCAATGGOCCCTACCTGGGGCCTCCCCAATCCCACTCTCTGCCTGATCTCCCAGACCAGCTGOGTTCCCCCCAGGTGGTOGGAGCCT CCA of the AIC2B cDNA. The AIC2B C 0 T iS shown the 2154 sequence in upper GCACTGCCAGGCTCCTTGGAGTACATGTGTCTGGCCCCCTGGAGGTCAAGTGCAACTOGTcTCCATTGTCCCAGGTGATGGGGCACGGCCAGGCTATGGATGTGCAGTGTGCMTCCAOCCTGC C A rorowssadteand the substitutions,ub iuindl-dele- 2274 GAGACCTCAGOGAGCCCTTCTGTGGAOCCAGGCAGACCCTCCAGTTGAGCTGAGCATGGAGGACAGGAGGCACGACAACCCAGTGACTCTGCCCATAAGCTCTOGGGGCCCTGAG tions, and insertions in the AIC2A A C 0 A A A T sequence are shown in the lower 2394 GGCAGTATGATGGCCTCTGATTATGTCACTCCTOGAGATCCGGTGCTCACTCTGCCCACAGGGCCCCTGTCTACCTCTCTOGGCCCCCTCTCTAGM'ffGCCCTCAGCCCMAOCCCCAGT rows. The initiation codon and the 2514 CTCTGTCTTAGCcTGCCCAGGGTCCCCTCTGGAAGCCCAGCTCTAGGGCCACCAGcGGMGAGACTATGTGGAGCTGCCTCCATGTGAOCCAcGOCTOGc TCCCCTCCAOCCAT termination codon are marked by A C >. The HindIII restriction site 2634 CCTGCTCCTCCTGTGCAAGCCAGCCCCCACAGTGATCCCAOGAGAGCCCAGOGAGGAAGTGGCCCCAGCATCCCCACATCCCGAAGGCCTCCTTGTTCTTCAOCAGGTTGGGGACTACTGC G used for the S1 protection assays 2754 TTCCTCCCTGGCCTOGGACCTGGCTCCCTCTCACCACACAGTAAGCCACCCTCTCCAAGTCTGTGTTCTGAGACTGAGGACCTAGTCCAGGACTTGTCTGTCMAAAATTTCCCTATCAG ~~~~isi indicated.niae.AA stretchsrthoof nucleo-ulo A 0 end A tides marked by * indicates the 2874 CCCATGCCCCAGGCCGCCAGCCATTCAGTmTMCATCCCTAGCTATCAGGACTACCTGTCCCTGCCCCCTTGGGACATAGCCAGTCTGGGAGGTGTGCTGAOTCTGTCTCCTCCCA region that oligonucleotide probes TC TT C ~ 2994 ATCTCACCAGCAGCCTGGCACCCCAGCCTGTGGTCCTCAGCCTGAGCATCACCACAGAAGCCTCTCTGAGTTCACACTCCTCCTTGCTCCCAGCCCTGACATGGM~TACCCCCACCTGT ~~~~~~~~~~were~~~~ee made.. for.o in the Southern T T CM TA analysis in Fig. 4. domain of the AIC2B protein has features similar to those of most of the changes which occur in the AIC2B protein are the IL-3 receptor such as proline-rich and serine-rich motifs conservative. However, several changes in the sequence (Fig. 3). Moreover, the secondary structure as predicted by potentially affect local secondary structure. For example, the the Choiu-Fasman method (17) is also well retained because addition of a helix-disrupting residue, glycine, at amino acid

1 V D 4 A L T 9 G. L C Y V A L V A L C I G |HQ V T E A EE T V P L K T L Q C Y N D Y T N H I I C SV A D T E D A q G L E E E R

B1 I N H T - L Y H Q L E K K q P V S C E L S E K L H* S E C P S S H R C V P R R C V I P Y T R FS I T N E D YY S F R P D L D I S N C D N 1W2S D L C I q L H V PL A q N V q P P L P K NV S I S SS E D R F L L EVI S V S LG D A Q V S * L S S K D I E F E V A Y K H P DIH P G H S 18U9 L D SV E D A Y S L H T S K F V N F E P K L F L PN S I Y A P R V R T R L Y P G S SL S C R P S R I S P E A V D S N L A S AV

240 S q P C D K A Q P q NL q C F F D C I q S L HC S I E V V T Q TT C S VS F C L F Y R P S P V A P E E K C S P V V K E P S A SA A II 310 P C A S VYT R Y H C SL P V P E P SA H S 4 Y T V S V K HL E q G K F I N S Y N H I q H E P P T L N L T K N R D S Y S Q- R Y I q 360 L H I E T q K H A Y S F I E H T F q V q Y KK K S D S V ED S K T E NL D RA H S H DL S q L E P D T S Y C A R V R V K FIG. 3. Amino acid sequence I P - K Y D E K C V N P of AIC2B. The amino acid se- 420 P I SN YD C I V S KV S E E Y T VK T D V V P I LV I V L I L V F L I L I L L L I L R F G C V S V Y R T Y R K I K E of the AIC2B is D E N T F A H R YCG quence protein shown in the upper rows and the 410 K I P N P SKSL L F q D C C K C L 9 P P C S H A A F A T K N P A L q C P q SR LLA E qG E S Y AH L E D NN V S P substitutions, insertions, and de- letions in the AIC2A protein are 540 L T I E D P N II R V P P S C P D T T P A A SS E S T E Q L P N V q V E C P T P - N R P R K q L P S F D F N C PYL C P shown in the lower rows. Amino D R I SS acid deletions are marked by 599 P q S HSL P D L P D q L C S P q V C G SL K P A L P G S L E Y N C L A P G G q V q L V P L S q V H G q G q A H D V q C and G P dashes. The signal sequence the transmembrane domain are B59 C S S L E T S C S P S V E P K E N P P V E L S H E E q E A R DN P V T L P I SS C C P E C S HV AS D YV T P C D P VL marked by an underline. The po- T VK H tential N-linked glycosylation 719 T L P T C P L S T S L C P S L C L P S A Q S P S L C L K L P RV P S C S P A L C P P C F E D Y V E L P P S V S q A A K S T sites are shown by +. The con- served cysteine residues are 779 P P C H P A P P V A S S P TV I P C E P R E E V C P A S P H P E C L L V L V G D YC F L P C L C P C S L S P H S KP R shown by *. Two homologous seg- ments in the external domain are 839 P S P S L C S E T E D L V Q D L S VKK F P Y Q P V P Q A P A I Q FF K S L K H q D Y L SL P P I DN S q S G KV C D L Y ------boxed. Downloaded by guest on September 23, 2021 5462 Biochemistry: Gorman et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1 2 3 FIG. 4. Southern hybridization with oli- a b gonucleotide probes specific to AIC2A and 2 3 4 12L 5 .+ n 15 kb AIC2B. EcoRI-digested mouse gemonic 4 1 kb DNA was separated on an agarose gel and free transferred to a nylon membrane filter probe -- (Schleicher & Schuell). The oligonucleotide A - probes used are as follows: lane 1, AIC2A- .838 specific 21-mer (CTGGGTCGGCTAGAAG- GAATT) (see Fig. 2); lane 2, AIC2B-specific 20-mer (CCTAGGTCTGTTAGGAGTTG) (see Fig. 2); and lane 3, a mixture of two probes. 350 - 40 &A* -_ 301 may lead to a loss of a-helical structure. Between residues 367 and 371 there are several amino acid changes B 270 that appear to increase the local hydrophobicity. In addition, other changes which may cause the loss of IL-3 binding 250 - include (i) charge switches at amino acid positions 388 and 430, (ii) addition of helix breakers at positions 214 and 222, and (iii) a switch from glycine at 396 to a charged aspartic residue. The relationship of the receptor structure to IL-3 FIG. 5. S1 nuclease protection assays using the AIC2A probe. binding remains to be determined. RNAs added were as follows: (a) Lane 1, IC2; lane 2, COS-7 cells Expression of the AIC2B cDNA in Fibroblasts. To test transfected with the AIC2A cDNA; lane 3, COS-7 cells transfected whether the AIC2B protein binds any known cytokines, we with AIC2B cDNA; lane 4, tRNA. (b) Lane 1, IC2; lane 2, PT18; lane expressed the AIC2-5 and AIC2-522 cDNAs in COS-7 cells 3, FDCP2; lane 4, FDCP2(-); lane 5, MC/9; lane 6, tRNA. A and B by using the expression vector pCEV4 (4), and we examined indicate AIC2A and AIC2B RNA protected bands. the cells for specific 251I-labeled cytokine binding. Although COS-7 cells expressed the AIC2B protein, which was con- AIC2B genes may be coregulated. Neither the A- nor the firmed by staining with anti-Aic2, they did not bind 251I-IL-3 B-type AIC2 RNA was detectable in T cells (HT-2, D10, at a concentration of ligand (2.8 nM) shown to give appre- HDK), stromal cells (ALC8, 30E, and 30R.7), or a fibroblast ciable binding to the AIC2A protein (Fig. 1B). Calculations cell line (L). However, a macrophage cell line (P388) and indicate that if AIC2B itself binds IL-3, it must do so with an CD5+ B-cell lines (CH12, CH44), which are IL-3-nonre- extremely low affinity, not accurately measurable with the sponsive in proliferation assays, expressed a relatively high given techniques. If AIC2B is involved in the formation of a level of the AIC2A and AIC2B RNA (Table 1). In all cells high-affinity IL-3 receptor complex by interaction with tested the level of the AIC2B RNA was always higher than AIC2A, a known IL-3-binding protein, then the affinity of that of AIC2A. IL-3 binding would be altered upon coexpression of AIC2A and AIC2B in the same cells. However, no such change was DISCUSSION noted (data not shown). We have therefore concluded that AIC2B is not a binding protein for IL-3, and we have Further characterization of the cDNAs cloned with the examined its ability to bind other cytokines. However, IL-2 anti-Aic2 antibody has revealed that two distinct yet homol- (7.6 nM), IL-4 (0.7 nM), GM-CSF (4.5 nM), erythropoietin ogous cDNAs are present in the cDNA library of the mouse (12.7 nM), and IL-9 (p40) (4.1 nM) at the indicated concen- mast cell line MC/9. The differences in the nucleotide trations all failed to show any significant binding (data not sequences are dispersed throughout the entire coding region, shown). and oligonucleotide probes specific to each cDNA hybridized Expression of the Two Types of the AIC2 mRNA in Various to two different genomic fragments, indicating the existence Cells. Northern blotting of MC/9 RNA using the AIC2-2 of two distinct genes. Moreover, our isolation of two distinct cDNA as a probe showed only one broad band at about 4.5 genomic clones has proved the existence of two separate kb, although two types of cDNA were isolated from MC/9 genes (unpublished work). (4). This result indicates that the sizes of the AIC2A and Several genes with extensive homology to known receptor AIC2B mRNAs are very similar and cannot be distinguished genes have been isolated: the c-erbB2 protooncogene (neu) by Northern blotting. To examine expression of the AIC2A and c-erbB3 are homologues of the epidermal growth factor and AIC2B genes we performed S1 nuclease protection receptor gene and the c- protooncogene has a stucture assays using RNA prepared from COS-7 cells transfected similar to the structures of the genes for the platelet-derived with either the AIC2A or AIC2B cDNA. An end-labeled growth factor receptor and the macrophage colony- AIC2A cDNA probe was protected by RNA prepared from stimulating factor receptor. The c-erbB2 gene was shown to the AIC2A cDNA-transfected COS-7 cells, but it was only be activated as a transforming gene by a point mutation (18), partially protected by the RNA of AIC2B-transfected COS-7 and gene amplification of c-erbB2 was observed in some cells because of the base substitutions (Figs. 2 and 5). tumors (19), suggesting a functional importance for the c- Conversely, when the end-labeled AIC2B cDNA probe was erbB2 gene. The c-kit protooncogene is included in the mouse used, AIC2A RNA only partially protected the probe, W locus (20), and mutations in this locus cause a defect in whereas the larger DNA probe was protected by the AIC2B embryonic development and hemopoiesis (21), indicating an RNA (data not shown). Under the same conditions using the important function in development. However, in neither case AIC2A probe, RNA from the IL-3-dependent cell lines (IC2, has the ligand for the putative receptor been identified. The MC/9, PT18, FDCP2, NFS60) generated two types of pro- AIC2B gene is an example of such a homologue in the tected cDNA probe, indicating that both genes are expressed cytokine receptor family. in these cells (Fig. 5). Interestingly, an IL-3-nonresponsive The AIC2B protein has all the structural features common variant, FDCP2(-), derived from the IL-3-dependent FDCP2 to the external domains of the cytokine receptors such as the cell line, had no detectable IL-3 binding (ref. 27; unpublished IL-2 receptor /3 chain (7), IL-3 (4), IL-4 (8, 9), IL-6 (10), results) and did not express a detectable level of either GM-CSF (11), and erythropoietin (12) receptors. These ele- transcript (Fig. 5). This result suggests that the AIC2A and ments include four conserved cysteines and the unique Downloaded by guest on September 23, 2021 Biochemistry: Gorman et al. Proc. Natl. Acad. Sci. USA 87 (1990) 5463 Table 1. Expression of the AIC2 RNA involved in the formation ofthe high-affinity IL-3 receptor by Band intensity itself. Alternatively, the AIC2B gene might have been created by gene duplication of the AIC2A and have lost its function Cell line Cell type AIC2A AIC2B during evolution. Further studies are required for discrimi- IC2 Mast ++ nation between these possibilities. Coexpression of the AIC2A and AIC2B RNA in various MC/9 Mast +++ PT18 Mast ++ cells suggests the possibility that expression of these two +++ genes is coordinately regulated under the same mechanism. FDCP2 Myeloid +++ +++ However, some cells, such as P388, CH12, and CH44, FDCP2(-) Myeloid + NFS60 Myeloid expressed the AIC2 RNA abundantly (Table 1) yet expressed Ml Macrophage ++ the Aic2 antigen poorly (unpublished results). There may be P388 Macrophage ++ ++ some post-transcriptional regulatory mechanism for the + expression of the Aic2 antigen. During the screening of the J174 Macrophage + B5B3C4 Pre-B AIC2 cDNA by hybridization, we found several minor forms ofcDNAs in both AIC2A and AIC2B. Since it is common that BCL-1 B + CH12 B multiple forms ofmRNA are produced by differential splicing CH44 B in various cell surface proteins such as CD45 (25), CD8 (26), CH32 B and IL-4 receptor (8), AIC2 may also have such a mechanism K23Tr T to produce multiple forms of mRNA. It would be interesting HT2 T to study how the expression of the AIC2 genes is regulated. D10 T ALC8 Stromal We are grateful to Drs. T. Yokota, K. Hayashida, K. Maruyama, 30E Stromal and H.-M. Wang for helpful discussion and F. Vega for synthesizing 30R.7 Stromal oligonucleotides. We also thank Drs. F. Lee, A. O'Garra, C. Martens, and J. Jackson for providing RNA samples. DNAX Re- L Fibroblast search Institute of Molecular and Cellular Biology is supported by RNAs prepared from various cells were used to evaluate the Schering-Plough Corporation. expression of the AIC2 RNA by using the S1 nuclease protection assay. Intensity of the protected bands on the autoradiogram was 1. Schrader, J. W. (1986) Annu. Rev. Immunol. 4, 205-230. visually classified. 2. Barton, B. E. & Mayer, R. (1989) J. Immunol. 143, 3211-3216. 3. Miyajima, A., Miyatake, S., Schreurs, J., De Vries, J., Arai, N., Yokota, T. & Arai, K. (1988) FASEB J. 2, 2462-2473. WSXWS sequence which is located close to the transmem- 4. Itoh, N., Yonehara, S., Schreurs, J., Gorman, D. M., Maruyama, K., brane domain. It is likely that AIC2B is a receptor for a Ishii, A., Yahara, I., Arai, K. & Miyajima, A. (1990) Science 247, known or an unknown cytokine. The amino acid sequence 324-327. identity of AIC2B to the IL-3 receptor (AIC2A) is unusually 5. Yonehara, S., Ishi, A., Yonehara, M., Koyasu, S., Miyajima, A., Schreurs, J., Arai, K. & Yahara, I. (1990) Int. Immunol. 2, 143-150. high (91%) compared to a 30% identity in the external 6. Schreurs, J., Arai, K. & Miyajima, A. (1990) Growth Factors 2, 221-234. domains ofthe a and / platelet-derived growth factor (PDGF) 7. Hatakeyama, M., Tsudo, M., Minamoto, S., Kono, T., Doi, T., Miyata, receptors (22), both ofwhich bind the BB form ofPDGF (23). T, Miyasaka, M. & Taniguchi, T. (1989) Science 244, 551-556. The receptors for growth hormone and also have a 8. Mosley, B., Beckmann, P., March, C. J., Idzerda, R. L., Gimpel, S. D., VandenBos, T., Friend, D., Alpert, A., Anderson, D., Jackson, J., structural motif similar to the cytokine receptor family and Wignall, J. M., Smith, C., Gallis, B., Sims, J. E., Urdal, D., Widmer, have a 33% identity (24). Although the similarity between M. B., Cosman, D. & Park, L. S. (1989) Cell 59, 335-348. these two hormones is weak, they compete with each other 9. Harada, N., Castle, B. E., Gorman, D. M., Itoh, N., Schreurs, J., for binding to both receptors (24). Therefore a specific ligand Barrett, R. L., Howard, M. & Miyajima, A. (1990) Proc. Natl. Acad. Sci. USA 87, 857-861. for the AIC2B protein, if it is present, may have a similar 10. Yamasaki, K., Taga, T., Hirata, Y., Yawata, H., Kawanishi, Y., Seed, structure to IL-3. The external domains ofAIC2A and AIC2B B., Taniguchi, T., Hirano, T. & Kishimoto, T. (1988) Science 241, can be divided into two relatively homologous segments, 825-828. each containing the conserved motifs ofthe cytokine receptor 11. Gearing, D. P., King, J. A., Gough, N. M. & Nicola, N. A. (1989) EMBO J. 8, 3667-3676. gene family. Preliminary results indicate that the second 12. D'Andrea, A. D., Lodish, H. F. & Wong, G. G. (1989) Cell 57, 277-285. segment of the AIC2A protein is important for IL-3 binding 13. Bazan, J. F. (1989) Biochem. Biophys. Res. Commun. 164, 788-795. (H.-M. Wang and A.M., unpublished data), suggesting an 14. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, W. J. alternative possibility, that the first segment is responsible for (1978) Biochemistry 18, 5294-5299. 15. Berk, A. J. & Sharp, P. A. (1978) Proc. Natl. Acad. Sci. USA 75, the binding to the unidentified ligand. Since the cytoplasmic 1274-1278. domains of both proteins are also well conserved, a putative 16. Seed, B. & Aruffo, A. (1987) Proc. Natl. Acad. Sci. USA 84, 3365-3369. ligand for AIC2B may induce a biological response similar to 17. Chou, P. & Fasman, G. (1978) Adv. Enzymol. 47, 45-148. IL-3. 18. Bargmann, C. I., Hung, M.-C. & Weinberg, R. A. (1986) Cell 45, 649-657. The AIC2B protein did not bind any cytokine (IL-2, IL-3, 19. Semba, K., Kamata, N., Toyoshima, K. & Yamamoto, T. (1985) Proc. IL-4, GM-CSF, erythropoietin, or IL-9) at concentrations Natl. Acad. Sci. USA 82, 6497-6501. between 1 and 10 nM when expressed in COS-7 cells. 20. Geissler, E. N., Ryan, M. A. & Housman, D. E. (1988) Cell 55, 185-192. However, this experiment still does not exclude the possi- 21. Russell, E. S. (1979) Adv. Genet. 20, 357-459. 22. Claesson-Welsh, L., Eriksson, A., Westermark, B. & Heldin, C.-H. bility that AIC2B is a component of one of these cytokine (1989) Proc. Natl. Acad. Sci. USA 86, 4917-4921. receptors. The observation that the IL-2 receptor /3 chain did 23. Heldin, C.-H., Backstrom, G., Ostman, A., Hammacher, A., not bind IL-2 when expressed in COS-7 cells but bound IL-2 Ronnstrand, L., Rubin, K., Nister, M. & Westermark, B. (1988) EMBO when coexpressed with the IL-2 receptor a chain (7) led us J. 7, 1387-1393. 24. Boutin, J. M., Jolicoeur, C., Okamura, H., Gagnon, J., Edery, M., to speculate that AIC2B may be a component of a cytokine Shirota, M., Banville, D., Dusanter-Fourt, I., Djiane, J. & Kelly, P. A. receptor complex. Because these two genes are coexpressed (1988) Cell 53, 69-77. in many cases, the possibility that the AIC2B protein is a 25. Streuli, M., Hall, L. R., Saga, Y., Schlossman, S. F. & Saito, H. (1987) component of the IL-3 receptor was examined. However, as J. Exp. Med. 166, 1548-1566. 26. Giblin, P., Ledbetter, J. A. & Kavathas, P. (1989) Proc. Natl. Acad. Sci. described above, cotransfection of the AIC2A and AIC2B USA 86, 998-1002. cDNAs into COS-7 fibroblasts did not change the affinity of 27. Sugawara, M., Hattori, C., Tezuka, E., Tamura, S. & Ohta, Y. (1988) J. IL-3 binding, suggesting that the AIC2B protein is not Immunol. 140, 526-530. Downloaded by guest on September 23, 2021